1484
Comparing productivity of pure and mixed
Douglas-fir and western hemlock plantations in
the Pacific Northwest
M.M. Amoroso and E.C. Turnblom
Abstract: We studied pure and 50/50 mixtures of Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) and western
hemlock (Tsuga heterophylla (Raf.) Sarg.) plantations to compare attained total yields between mixed-species stands as
opposed to monocultures of equal densities. Whether overall stand density influences this outcome has not been adequately
investigated, and to address this we included three density levels (494, 1111, and 1729 trees/ha) in the analysis. At age
12, as components of the mixed stands, Douglas-fir exhibited greater height, diameter, and individual-tree volume than
western hemlock at all densities. At 494 and 1111 trees/ha the monocultures had a higher volume per hectare than the
mixed stand, but at 1729 trees/ha the mixed stand appeared to be just as productive as the pure stands. The increase in
productivity by the mixture at high densities seems to have resulted from the partial stratification observed and most
likely also from better use of the site resources. Because of this, less interspecific competition was probably experienced
in the mixed stand than intraspecific competition in the pure stands. This study shows the important role density plays
in the productivity of mixed stands and thus in comparing mixed and pure stands.
Résumé : Les auteurs ont étudié des peuplements purs et mélangés (50/50) de douglas de Menzies (Pseudotsuga menziesii
(Mirb.) Franco) et de pruche de l’Ouest (Tsuga heterophylla (Raf.) Sarg.) en plantation pour comparer les rendements
totaux obtenus en peuplements mélangés comparativement à des monocultures de même densité. Ils ont inclus dans
l’analyse trois niveaux de densité (494, 1111 et 1729 tiges/ha) étant donné que l’effet de la densité globale du peuplement sur le rendement n’a pas été adéquatement étudié. À l’âge de 12 ans, en peuplements mélangés, la hauteur, le
diamètre et le volume des tiges individuelles de douglas de Menzies étaient supérieurs à ceux de la pruche de l’Ouest
peu importe la densité. Avec 494 et 1111 tiges/ha, les monocultures avaient un volume à l’hectare plus élevé que le
peuplement mélangé mais avec 1729 tiges/ha le peuplement mélangé semblait tout aussi productif que les peuplements
purs. L’augmentation de productivité dans le mélange à haute densité semble être le résultat de la stratification partielle
qui a été observée et fort probablement aussi d’une meilleure utilisation des ressources du site. À cause de cela, il y
avait probablement moins de compétition interspécifique dans le peuplement mélangé que de compétition intraspécifique
dans les peuplements purs. Cette étude montre l’importance du rôle que joue la densité sur la productivité des peuplements mélangés et, par conséquent, dans la comparaison des peuplements purs et mélangés.
[Traduit par la Rédaction]
Amoroso and Turnblom
Introduction
When timber production is the primary objective of management, there is a clear tendency to favor monocultures of
the most productive species. This is mainly because of the
simpler management of monocultures (easier and cheaper
establishment, planning and marketing), but also, though no
less important, because less is known about planted mixed
stands and the interactions between species. In contrast, when
mixed-species stands are favored, the objectives usually
Received 12 September 2005. Accepted 2 February 2006.
Published on the NRC Research Press Web site at
http://cjfr.nrc.ca on 17 May 2006.
M.M. Amoroso1,2 and E.C. Turnblom. College of Forest
Resources, University of Washington, Box 352100, Seattle,
WA 98195-2100, USA.
1
2
Corresponding author (e-mail: amoroso@interchange.ubc.ca).
Present address: Department of Forest Sciences, The
University of British Columbia, 3219-2424 Main Mall,
Vancouver, BC V6T 1Z4, Canada.
Can. J. For. Res. 36: 1484–1496 (2006)
1496
include wildlife conservation, aesthetics, resistance to wind
damage, risk reduction or compensatory growth, and protection
from disease and insect outbreaks. A sacrifice in productivity,
compared to the most productive species, is usually assumed
to occur as a consequence of the use of mixed-species stands
(Kelty 1992). This research examines the validity of this
assumption, including the possibility of achieving comparable
or greater total yields when using mixed-species planted
stands as opposed to monocultures of equal densities.
Ecological theory suggests that species in a mixture may
exploit resources of a site more completely and efficiently
than a single species would be able to do, leading to greater
overall productivity (Vandermeer 1989). Even though this
has been observed in many situations, it is not always likely
to happen. To achieve greater productivity in mixed stands,
the species constituting the stands need to show differences
in their requirements (niches) and the way they use site
resources and (or) positively affect the growth of each other
(Vandermeer 1989). This concept of niche separation implies
that if two species are too similar in their requirements they
would eventually compete intensely to exclude the other, but
doi:10.1139/X06-042
© 2006 NRC Canada
Amoroso and Turnblom
if competition is sufficiently weak, the two species may
coexist (Harper 1977).
The principal mechanism that has been used in forestry to
increase production in mixed stands is to increase nutrient
availability (Kelty and Cameron 1995) primarily through the
introduction of nitrogen-fixing species in mixtures (Binkley
1983, 1992; DeBell et al. 1997; Khanna 1997; Bauhus et al.
2000; Balieiro et al. 2002). Furthermore, photosynthetic
efficiency of foliage, differences in height growth patterns,
form, phenology, and root structure have been suggested as
possible causes leading to a mixed stand having overall productivity greater than a monoculture (Kelty 1992).The mixed
stands may experience less intense interspecific than
intraspecific light competition as a consequence of the
differences in shade tolerance among species. Such a stratified
canopy would, in theory, maximize the use of light because
of increased light interception and light-use efficiency (Kelty
1992), leading to greater total productivity than in pure stands
(Smith et al. 1997). This type of response has been found in
studies by Assmann (1970), Wierman and Oliver (1979),
Kelty (1989), Brown (1992), Montagnini et al. (1995, 2000),
DeBell et al. (1997), Man and Lieffers (1999), and Garber
and Maguire (2004).
Previous research has found that Douglas-fir (Pseudotsuga
menziesii (Mirb.) Franco) and western hemlock (Tsuga
heterophylla (Raf.) Sarg.) differ in their shade tolerance
(Lewis et al. 2000) and height growth. When growing together,
Douglas-fir, a shade-intolerant species, tends to occupy the
upper part of the canopy, while shade-tolerant western hemlock develops in the understory (Oliver and Larson 1996).
This stratified canopy structure in even-aged stands also tends
to develop naturally, because sun-adapted species generally
have greater rates of juvenile height growth than shadetolerant species when growing together (Kelty 1992; Oliver
and Larson 1996; Smith et al. 1997). There is abundant
evidence of differing height growth patterns, and consequent
stratification, for these two species in natural stands (King
1958; Scholz and Smith 1975; Wierman and Oliver 1979;
Oliver and Larson 1996). Wierman and Oliver (1979) found
that in even-aged mixed natural stands of these two species,
Douglas-fir dominated and suppressed competing western
hemlock trees by becoming significantly taller after about
20 years and that basal area per acre was on average greater
in the mixed stands than in the pure stands; the authors also
concluded that mixed stands appeared to yield greater volume
per unit area than pure stands.
In accordance with what previous studies have shown,
these two species are likely to exhibit “ecological combining
ability” (Harper 1977) and to have greater productivity in a
mixed stand than in pure stands of their respective components. However, all the previous research on Douglas-fir and
western hemlock mixtures has only examined natural stands.
Questions remain as to whether the patterns of stratification
and growth of these species observed in natural stands also
occur in plantations and how stand growth, yield, and structural development depend on initial stand density.
The objectives of this study were to assess differences in
growth and productivity between Douglas-fir and western
hemlock growing in both pure and mixed plantations across
a range of planting densities. This study, consisting of pure
Douglas-fir, pure western hemlock, and a 50/50 mixture of
1485
these two species combined with three density levels and
replicated at two sites, provides a well-designed field experiment testing the effects of widely varying spacing and species
composition.
Methodology
Experimental sites
This study was conducted at two sites (Brittain Creek and
Forks) located on the Olympic Peninsula in the state of
Washington, USA. The Brittain Creek site is located in Gray’s
Harbor County (47°13′N and 123°52′W), while the Forks
site is located in Clallam County (48°02′N and 124°23′W).
Brittain Creek has a mean annual precipitation of approximately 2902 mm and a mean annual temperature of 9.7 °C
(Western Regional Climate Center 2003). The site is located
at an elevation of 110 m, and the site index is 38 m (125 ft)
at 50 years, corresponding to a site class of II (King 1966).
The site has a 1%–15% slope, and the soil, a Willaby silt
loam, consists of very deep, moderately well-drained soils
from the Willaby series formed in glacial drift (Pringle 1986).
Forks has a mean annual precipitation of approximately
3047 mm and a mean annual temperature of 9.8 °C (Western
Regional Climate Center 2003). This site is located at an
elevation of 122 m, and the site index is 37 m (120 ft) at
50 years, corresponding to a site class of II (King 1966).
The site has a 0%–15% slope, and the soils are classified as
part of the Klone–Ozette–Tealwhit complex from the Klone
series and characterized as very deep, well-drained soils
formed in poorly sorted glacial outwash (Halloin 1987).
These two study sites are part of a number of research
installations established by the Stand Management Cooperative
(SMC) across the Pacific Northwest with the objective of
designing, establishing, and maintaining a regional program
of integrated research on various aspects of intensive stand
management (Maguire et al. 1991). The two installations
chosen for this study were established and planted in 1990
with Douglas-fir 2–1 stock and western hemlock 1–1 stock.
Experimental design
At each of the experimental sites, a block containing three
installations (sub-blocks) was established. Species composition,
consisting of pure Douglas-fir, pure western hemlock, and a
50/50 mix of these two species, was then randomly assigned
to the three installations. Within each of these installations,
six areas (experimental units) of at least 3 acres (1 acre =
0.404 686 ha) in size were delineated, where one of six
initial desired planting densities (245, 479, 748, 1074, 1682,
and 2989 trees/ha) was randomly assigned to each experimental unit, giving the experiment a randomized complete
block split-plot design. In each of these experimental units, a
0.405 ha plot was located. Each plot includes a buffer strip
and a measurement sample plot (MSP), the size of which
varies depending on the target spacing in each spacing block.
The MSP dimensions, shown in Table 1, were selected as
multiples of the target spacing for each block (Maguire et al.
1991). Thus, there are 36 MSPs, and it should be emphasized that the MSPs are the experimental units; therefore,
they are to be treated as the independent observations,
because they represent the stage at which randomization of
planting densities occurred (Oehlert 2000). Kelty and Cameron
© 2006 NRC Canada
1486
Can. J. For. Res. Vol. 36, 2006
Table 1. Sample plot size and characteristics by targeted planting
spacing.
Spacing
(m)
6.4
4.6
3.7
3.0
2.4
1.8
Trees/ha
245
479
748
1074
1682
2989
No. of
rows
7
8
9
10
13
16
Plot side
(m)
Plot size
(ha)
Trees
per plot
44.8
36.6
32.9
30.5
31.7
29.3
0.20
0.13
0.11
0.09
0.10
0.09
121
158
200
247
417
632
(1995) discuss other individual-tree study designs, which are
for many reasons not appropriate for the type of analyses
presented here.
Measurements and field data collection
Basal diameter and diameter at breast height (DBH) were
measured on all trees. Basal diameter was measured 18 cm
above the ground line using calipers until trees reached breast
height, while DBH measurements were done by caliper or
diameter tape (to the nearest 0.25 cm). Total height and
height to the base of the live crown were measured with a
laser hypsometer to the nearest 0.03 m on a sample of 42
trees drawn from the diameter distribution, including the
smallest and the largest trees on each plot. Crown width
(average of two crown widths: east–west and north–south,
measured to the nearest 0.03 m) was also measured on the
same subsample of trees. Four measurements taken on a
2-year cycle and an additional measurement on a 4-year cycle
were available for analysis.
Data analysis
Trees without measured heights in each plot were given a
height estimate from a nonlinear height–diameter equation fit
by species to all 42 height trees measured in the plot at any
given time. We used a functional form recommended by
Martin and Flewelling (1998), because they found it to be
properly parameterized and unbiased when tested on inventory
data sets from western Washington. The following individualtree attributes were derived from diameter and height: basal
area, height/diameter (h/d) ratio, crown ratio, and tree volume.
Douglas-fir tree volumes were calculated using the method of
Bruce and de Mars (1974), and western hemlock volumes were
estimated using the method of Flewelling and Raynes (1993).
Mean values for each plot and measurement number were
calculated for all the variables described previously, as well
as for basal area per hectare, volume per hectare, and quadratic mean diameter.
The targeted planting densities were met with variable
success in the field. The density on a few of the plots varied
so widely from their target density that they actually matched
the next higher or lower planting density. Further, measurements on the two research sites, Forks and Brittain Creek,
began 1 year apart. So, comparisons among plots with the
same density at the six fixed initial planting densities at a
specific age either were not possible or would be imprecise
at best. Therefore, to enhance precision of the comparisons
among stands, regression analysis was used, in which models
for the variables under analysis were built to predict values
at any given year and density within the range of final
densities and the time frame of the study. The two primary
independent variables in the models were trees per hectare
(the actual trees per hectare at each measurement) and age
(time since planting). In addition, because one of the objectives of this study was to assess the behavior of each species
in both pure and mixed stands, two categorical variables
were included in the model: one corresponding to “species”,
to differentiate between Douglas-fir and western hemlock in
the pure stands, and another one corresponding to “component”,
to differentiate between the two species when growing
within the mixtures. The full quadratic response surface
models included the independent and categorical variables
mentioned previously and all possible interactions among
them. A quadratic response surface model was chosen because
it is a well-known standard analysis technique for welldesigned experiments and is relatively easy to interpret (Oehlert
2000). It allows dependent variables to exhibit level, linear,
and (or) curvilinear responses to both the experimentally
manipulated variables of species composition and density
and to the covariate age. It also allows for testing of interaction effects between all independent variables. Higher order
response surfaces, such as the cubic, which are capable of
identifying more complex nonlinear responses, were not
considered necessary, given the young age of the plantations. The general model form is presented in the Appendix.
Separate quadratic response surface models were fit to
each dependent variable using multiple regression analysis.
Residual plots indicated that all response variables needed to
be transformed because of non-constant variance and nonnormality. Square-root (for DBH) and logarithmic transformations (for remaining variables) were sufficient to stabilize
variance and improve normality in all cases. The final models
for all response variables contained only those predictor
variables with p ≤ 0.05. Possible autocorrelation effects with
the use of age as a covariate were not assessed. Rather, we
relied on the analytical heuristic that when the number of
independent observations (36 plots in this case) exceeds the
number of repeat measurements (5 in this case), the effects
of autocorrelation are very likely to be minimal (Crowder
and Hand 1996).
Utilizing these models, values for three chosen density
levels (494, 1111, and 1729 trees/ha) were initially predicted
at three ages (4, 8, and 12 years). The low- and high-density
levels correspond to the range of actual density values
within which all three stand types were represented; the situation is analogous for the three ages. All comparisons made
in the results were conducted by computing 95% confidence
intervals on the predicted responses at the chosen levels of
the independent variables, taking any nonoverlapping intervals
as sufficient evidence of statistical significance. Combined
type I error rates were controlled using Bonferroni’s procedure, so that all the confidence intervals constructed for any
given response variable (height, DBH, volume, etc.) hold
simultaneously, meaning confidence coverage is at least 95%
for any single predicted response (Oehlert 2000). Statistical
analyses were made using the SAS package for Windows,
version 8 (SAS Institute Inc. 1999–2001).
The yield of pure and mixed stands is usually compared
on a relative basis; thus, the effects of combining the two
species were evaluated by comparing the yield of each species
© 2006 NRC Canada
Amoroso and Turnblom
1487
Table 2. Mean stand characteristics predicted at age 12 by species composition and the three density
levels chosen for analysis.
Density
(trees/ha)
494
1111
1729
Composition
Height
(m)
DBH
(cm)
Douglas-fir
Western hemlock
Mixture
Douglas-fir
Western hemlock
Mixture
Douglas-fir
Western hemlock
Mixture
10.2
8.9
8.9
10.1
9.1
9.1
10.0
9.8
9.8
14.94
12.67
13.21
13.29
12.46
12.04
12.30
11.70
11.77
Basal area
(m2/ha)
9.8
5.7
6.8
19.1
16.8
14.5
24.4
18.6
24.6
Volume
(m3/ha)
39
26
27
78
82
60
104
94
109
Note: The density levels chosen for the analyses are well within the range of the six targeted planting densities displayed in Table 1.
Fig. 1. Mean total height by treatment at age 12 for three density levels. Bars represent 95% confidence intervals. Note: The
trend line for the mixed stand lies on top of the line for pure
hemlock.
< 1 indicate an antagonistic or competitive relationship
between the species in the mixture. In our case, the assumption
that 50/50 mixtures of Douglas-fir and western hemlock
grow independently would result in each species having an
expected RY = 0.5 and an RYT = 1.0.
12
DF
WH
MIX
df/MIX
wh/MIX
Results
Mean Height (m)
11
10
9
8
7
6
400
600
800
1000
1200
1400
1600
1800
Density (trees/ha)
in the mixture with its yield in monoculture as per Harper
(1977). Yield variables for this analysis were basal area and
volume per hectare. The relative yield (RY) of each species
and the relative yield total (RYT) were calculated as
RYDouglas − fir =
Yield of Douglas-fir in mixture
Yield of Douglas-fir in monoculture
RYWestern hemlock =
Yield of western hemlock in mixture
Yield of western hemlock in monoculture
Relative yield total (RYT) = RYDouglas − fir + RYWestern hemlock
If both species use resources in identical ways, and hence
compete for these resources, the expected RY of each species
will be equivalent to its proportional contribution in the
mixture, and an expected RYT = 1. An RYT > 1 indicates
either niche separation or the existence of some beneficial
relationship between species, producing a potential productivity gain for the mixture. On the other hand, values of RYT
Mean stand characteristics at age 12 for the two pure
stands and the mixed stand at the three chosen densities are
summarized in Table 2. At this age, as expected, diameter
had an inverse relationship with density in all situations.
Height, instead, was found to have different responses with
density depending on the stand composition. Both basal area
and volume per hectare increased with density, and there
were changes in terms of the most productive stand composition at different densities. Hereafter, pure Douglas-fir
stands will be denoted DF, pure western hemlock stands
WH, and mixed stands MIX. Douglas-fir and western hemlock as components of the mixed stands will be denoted
df/MIX and wh/MIX, respectively. More detailed and deeper
analyses for all the variables studied are presented in what
follows.
Patterns of height growth
Ninety-five percent of the total variation in the logarithm
of height was explained by the following predictor variables:
Douglas-fir component indicator (C1), the western hemlock
component indicator (C2), the Douglas-fir species indicator
(SP1), TPH, AGE, C1 × AGE, C2 × AGE, SP1 × AGE,
TPH × TPH, AGE × AGE, SP1 × AGE × AGE, and C2 ×
AGE × AGE, where TPH is trees per hectare (see Appendix
for full model form).
Height is known to be affected only at extreme ranges of
density; however, some differences were found to be significant at age 12 for the range of densities used in this study
(Fig. 1). It seems that in pure stands the height of Douglasfir is less affected than that of western hemlock over the
range of densities examined. Western hemlock trees growing
at a density of 1729 trees/ha were significantly taller than
hemlock at lower densities. Comparisons between the two
species show that Douglas-fir was significantly taller than
western hemlock by more than 1 m at 494 trees/ha; however,
differences in total mean heights for the two species were in© 2006 NRC Canada
1488
Can. J. For. Res. Vol. 36, 2006
Fig. 2. Mean total height over time by species for two density levels: 494 and 1729 trees/ha.
494 trees/ha
12
1729 trees/ha
12
DF
WH
10
wh/MIX
Mean Height (m)
Mean Height (m)
df/MIX
8
6
4
2
8
6
4
2
0
0
2
3
4
5
6
7
8
Age (years)
9
10
11
12
Table 3. Periodic annual increment in mean total height (m) for
Douglas-fir (DF) and western hemlock (WH) growing in pure and
in mixed stands for three growth periods.
2
4
494
1111
1729
0–4
DF
WH
df/MIX
wh/MIX
DF
WH
df/MIX
wh/MIX
DF
WH
df/MIX
wh/MIX
0.63
0.64
0.64
0.60
0.81
0.71
0.79
0.56
1.11
0.86
1.05
0.81
0.62
0.66
0.66
0.61
0.80
0.73
0.81
0.57
1.09
0.88
1.07
0.83
0.62
0.70
0.71
0.66
0.80
0.79
0.87
0.61
1.09
0.95
1.15
0.89
Note: df/MIX and wh/MIX denote Douglas-fir and western hemlock,
respectively, as components of the mixed stands.
DF
9
10
11
12
WH
MIX
df/MIX
1000
1200
1400
wh/MIX
14
13
12
11
10
9
8
7
400
significant at 1729 trees/ha. There were no significant height
differences at age 8 or age 4 for any density in the pure
stands.
When comparing the two species as components of the
mixed stands, results were different. Douglas-fir was on
average 2 m taller than western hemlock across densities.
These differences in the height growth pattern started at
around age 5 (Fig. 2) and were a consequence of the differences in height growth rate for the two species (Table 3);
right after age 4 Douglas-fir started growing faster in height
than western hemlock.
Compared with stands growing in a mixture, western
hemlock was on average 1 m taller growing in pure stands.
Douglas-fir, instead, had the same height both in pure and in
mixed stands when it grew at 494 and 1111 trees/ha but
became about 1 m taller at 1729 trees/ha. This change in the
growth pattern seems to be more evident beginning at age 8.
The overall trends through time of all the patterns described
previously can be visualized in the set of graphs at two
different densities (Fig. 2).
6
7
8
Age (years)
15
Mean Diameter (cm)
Species
5
17
16
Growth period (years)
8–12
3
Fig. 3. Mean diameter by treatment at age 12 for three density
levels. Bars represent 95% confidence intervals.
Density (trees/ha)
4–8
DF
WH
df/MIX
wh/MIX
10
600
800
1600
1800
Density (trees/ha)
Patterns of diameter growth
Ninety-four percent of the total variation in the squareroot of DBH was explained by the following predictor
variables: Douglas-fir component indicator (C1), the western
hemlock component indicator (C2), the Douglas-fir species
indicator (SP1), the western hemlock species indicator
(SP2), TPH, AGE, SP2 × TPH, C2 × AGE, TPH × AGE,
TPH × TPH, AGE × AGE, SP1 × TPH × TPH, and SP2 ×
TPH × TPH (see Appendix for full model form).
Density and species composition had a strong effect on
mean diameter at age 12. As was expected, diameter had an
inverse relationship with density, and it appears that the two
species responded to density differently (Fig. 3). Growing in
pure stands, Douglas-fir was about 2.5 cm greater when it
grew at 494 than at 1729 trees/ha. Western hemlock, instead,
did not show significant diameter differences across the three
density levels, though a slight trend is observed that qualitatively matches that for Douglas-fir. Differences between the
© 2006 NRC Canada
Amoroso and Turnblom
1489
Fig. 4. Mean diameter over time by species for two density levels: 494 and 1729 trees/ha.
494 trees/ha
1729 trees/ha
16
14
10
8
6
4
2
10
8
6
4
0
2
3
4
5
6
7
8
Age (years)
9
10 11
12
Table 4. Periodic annual increment in mean diameter (cm) for
Douglas-fir (DF) and western hemlock (WH) growing in pure
and in mixed stands for three growth periods.
Density (trees/ha)
Growth period (years)
Species
494
1111
1729
0–4
DF
WH
df/MIX
wh/MIX
DF
WH
df/MIX
wh/MIX
DF
WH
df/MIX
wh/MIX
1.8
1.6
1.8
1.3
1.4
1.2
1.4
1.0
0.6
0.4
0.6
0.3
1.6
1.5
1.6
1.1
1.3
1.2
1.3
0.9
0.5
0.5
0.6
0.3
1.4
1.4
1.5
1.0
1.2
1.1
1.3
0.9
0.5
0.5
0.7
0.3
8–12
12
2
0
4–8
DF
WH
df/MIX
wh/MIX
14
Mean Diameter (cm)
12
Mean Diameter (cm)
16
DF
WH
df/MIX
wh/MIX
Note: df/MIX and wh/MIX denote Douglas-fir and western hemlock,
respectively, as components of the mixed stands.
two single-species stands were significant at 494 trees/ha;
however, they became insignificant at 1729 trees/ha; it
seems that Douglas-fir is more affected in its potential
growth at higher densities than western hemlock. Mean
diameter of the mixture did not show significant differences
with that from western hemlock at any of the densities and
became statistically indistinguishable from Douglas-fir at the
high-density level. The described differences were already
established at age 8.
Comparisons between species as components of the mixed
stands produced different results than pure stand comparisons. At age 12 Douglas-fir had a DBH that was on average
4 cm more than that of western hemlock at all densities
(Fig. 3). Furthermore, these differences appeared early in
stand development and were already on the order of 3 cm at
age 8 (Fig. 4). Growth rates for the two species at different
ages support this pattern (Table 4).
Apparently some changes in the growth pattern of the
species occur when they grow in pure as opposed to mixed
stands. Western hemlock was 2.5 cm greater in DBH when it
2
3
4
5
6
7
8
Age (years)
9
10
11
12
grew in pure stands, and this was consistent across all densities.
Although these differences were not found to be significant
at the 95% confidence level, Douglas-fir became more than
1 cm in DBH greater when it grew at high densities in the
mixture. The overall trends through time for all the patterns
described previously can be visualized in the set of graphs at
the different densities (Fig. 4).
Height/diameter ratio
The height/diameter ratio (h/d) was compared among
species while growing alone and together at each of the density levels (Fig. 5). Even though some significant differences
in diameter and height growth were evident at age 12, the
h/d ratio at this age is very low and is not significantly
different among the treatments over the range of densities
studied. Western hemlock growing in the mixture seemed to
be developing a slightly greater h/d ratio than other stand
types at this stage.
Individual-tree volume growth
Ninety-four percent of the total variation in the logarithm
of mean tree volume was explained by the following predictor
variables: Douglas-fir component indicator (C1), the western
hemlock component indicator (C2), the Douglas-fir species
indicator (SP1), the western hemlock species indicator (SP2),
TPH, AGE, C2 × TPH, C2 × AGE, SP2 × AGE, TPH ×
AGE, TPH × TPH, AGE × AGE, and C2 × TPH × TPH (see
Appendix for full model form).
At the three density levels, Douglas-fir trees in pure stands
had significantly larger mean tree volume than western hemlock trees in pure stands. It is interesting to note that even
though at 1729 trees/ha both height and diameter were not
significantly different between the two species, Douglas-fir
mean tree volume was significantly greater than that of
western hemlock. Although the individual-tree volume for
both species in pure stands decreased significantly with density from 494 to 1111 trees/ha, there was no significant
change from 1111 to 1729 trees/ha (Fig. 6).
Differences between the two species became more evident
in the mixed stands. Douglas-fir as a component of the mixed
stands produced about 22% higher mean tree volumes at all
densities when compared with Douglas-fir growing in pure
© 2006 NRC Canada
1490
Can. J. For. Res. Vol. 36, 2006
Fig. 5. Height/diameter ratio by treatment at age 12 for three
density levels. Bars represent 95% confidence intervals.
Fig. 6. Mean tree volume by treatment at age 12 for three density
levels. Bars represent 95% confidence intervals.
70
120
WH
MIX
df/MIX
wh/MIX
DF
Mean Tree Volume (dm ³ )
Height/Diameter Ratio
DF
60
50
40
30
20
WH
df/MIX
wh/MIX
100
80
60
40
20
10
0
0
494
494
1111
1729
1111
1729
Density (trees/ha)
Density (trees/ha)
stands at the same total stand density. In contrast, western
hemlock in mixed stands had mean tree volume reduced by
50%, 43%, and 53% at 494, 1111, and 1729 trees/ha, respectively.
The individual-tree volume of the two species growing in
mixed stands was also compared with those of the species
growing in pure stands but at the same density of each
component in the mixture (i.e., half of the total stand density
of the mixture). Mean tree volume of Douglas-fir did not
exhibit significant differences. Western hemlock mean tree
volume, on the other hand, experienced a statistically significant decrease in the mixture of 55% and 60% at 865 trees/ha
(half of 1729 trees/ha) and 555 trees/ha (half of 1111 trees/ha),
respectively.
Per-hectare volume accumulation
Ninety-three percent of the total variation in the logarithm
of volume per hectare was explained by the following
predictor variables: Douglas-fir component indicator (C1),
the western hemlock component indicator (C2), the Douglasfir species indicator (SP1), the western hemlock species indicator (SP2), TPH, AGE, C2 × TPH, SP2 × TPH, C2 ×
AGE, SP2 × AGE, TPH × AGE, TPH × TPH, AGE × AGE,
C2 × TPH × TPH, SP1 × TPH × TPH, and SP2 × TPH ×
TPH (see Appendix for full model form).
At low densities, pure Douglas-fir had higher volume per
hectare at age 12 than the pure western hemlock and the
mixed stand (Fig. 7). At 1111 trees/ha both the Douglas-fir
and western hemlock monocultures had significantly higher
volume per hectare than the mixture. At 1729 trees/ha volume per hectare was statistically indistinguishable among the
three stands. These results are consistent with what was
found at this density for the height and diameter growth patterns where no significant differences were found among
stand types. Another interesting result was that at 1729
trees/ha, Douglas-fir, as a component of the mixture, resulted
in volume per hectare statistically indistinguishable from
that of the pure hemlock stand. Results at age 8 were similar
to those found for basal area at age 12, with the only difference that volumes in the western hemlock stands were lower
than those for Douglas-fir at all densities.
The percentage contribution of Douglas-fir to the mixture
total volume was higher across densities, ranging from 74%
to 78%. Western hemlock contributed the balance of 22%–26%.
It appears that at densities higher than 1500 trees/ha,
volume per hectare was no longer significantly different
among the three stands. This result was paralleled by an
increment in productivity of the Douglas-fir component in
the mixed stands, which at the density of 1729 trees/ha
accounted for 25% more volume than the same number of
Douglas-fir trees growing in pure stands.
Volume per hectare in the western hemlock stands did not
result in significantly different values at 1111 and 1729
trees/ha. To examine this more closely, volume per hectare
for intermediate densities was estimated (Fig. 8). With this
new resolution, volume per hectare for western hemlock was
no longer significantly affected by densities greater than
1200 trees/ha.
Relative yield analysis
The effects of combining two species in a mixture were
analyzed by comparing the yield of each species in mixture
with its yield in a pure stand (Harper 1977). Relative yield
(RY) and relative yield total (RYT) were calculated for basal
area and volume per hectare, but because the results were
similar, only those for volume are presented (Fig. 9). The
relative yield of western hemlock was less than 0.5 at all density levels. Douglas-fir, instead, had RY values substantially
greater than 0.5 at both 1111 and 1729 trees/ha. It is clear
that the mixture of the two species benefitted the Douglas-fir
at 1111 and 1729 trees/ha as judged by its volume yield.
Combined RYT was less than 0.8 for both 494 and 1111
trees/ha but, although not statistically different, was higher
than 1.0 at 1729 trees/ha. Thus, it appears that at the highest
density significant niche separation between these species
may exist. Even though this suggests it might be a potential
advantage for the mixture compared to the monocultures,
© 2006 NRC Canada
Amoroso and Turnblom
1491
Fig. 7. Volume per hectare by treatment at age 12 for three density
levels. Bars represent 95% confidence intervals.
Fig. 8. Volume per hectare by treatment at age 12 across all density
levels. Bars represent 95% confidence intervals.
120
120
DF
WH
MIX
df/MIX
wh/MIX
DF
MIX
100
Volume (m3/ha)
100
Volume (m3/ha)
WH
80
60
80
60
40
40
20
20
0
400
0
400
600
800
1000
1200
1400
1600
1800
Density (trees/ha)
absolute yield values should be compared to identify the
highest yielding stand (Kelty 1992). This comparison is
shown in Fig. 10, which combines the results found for
absolute and relative yield at the three proposed densities.
This shows that at 1729 trees/ha both the relative and the
absolute yield (volume per hectare) for the mixture were
statistically indistinguishable from that of the two monocultures.
Discussion
Height growth pattern and stratification
The relationship between juvenile height growth rates and
shade tolerance among species plays an important role in
determining development patterns of mixed stands
(Menalled et al. 1998). Stratified canopies in mixed stands
tend to develop naturally because shade-intolerant species
generally have greater rates of juvenile height growth than
shade-tolerant species, and shade-tolerant species are able to
survive in reduced light environments (Kelty 1992; Oliver
and Larson 1996; Smith et al. 1997). Douglas-fir and western
hemlock differ in their tolerance to shade (Lewis et al. 2000),
and different height growth patterns along with stratification
have been found for the two species in natural stands (King
1958; Scholz and Smith 1975; Oliver and Larson 1996).
Working with even-aged mixed natural stands of these two
species, Wierman and Oliver (1979) reconstructed the height
growth pattern and found that after about 20 years Douglasfir was significantly taller than western hemlock. However,
whether Douglas-fir would similarly outgrow western hemlock
in mixed plantations and this stratification pattern occurs
was not certain. This plantation study shows that by age 12
Douglas-fir has outgrown western hemlock by an average of
2 m across all densities. Since the Douglas-fir trees were
1 year older than the hemlock trees at planting, admittedly it
may have had some small advantage, but no data or evidence
is available to suggest an initial height difference at time of
600
800
1000
1200
1400
1600
1800
Density (trees/ha)
planting. Differences in height growth were quite small
starting around age 4, but were almost 1 m at age 8 (Fig. 2).
These height growth rates demonstrate that the difference
between the two species has been increasing over the period
of time measured, and though this trend is expected to continue in the near future, western hemlock may later catch up
with Douglas-fir (Larson 1986). Even though the evidence
of stratification at this point is partial, the increasing juvenile
height growth observed and the height differences already
established for the two species in the mixed plantations support
the conclusion that stratification, even though not present
yet, is developing.
It has been suggested that heights of the upper canopy in a
mixed stand are expected to reach the same height as would
be achieved in the pure stands and that height is not sensitive
to stand density except at extremely wide and narrow spacing
(Oliver and Larson 1996). However, the present study found
that at the highest density (1729 trees/ha), Douglas-fir was
taller in the mixture than in the pure stands. These results
are consistent with those found by other authors working
with mixed planted stands (Menalled et al. 1998; Garber and
Maguire 2004); nevertheless, the height increment of dominant
trees in mixed stands is not expected to result as a consequence of a reduction in crown competition due to stratification
(Menalled et al. 1998).
Interspecific and intraspecific competition
It has been proposed that in some situations a more efficient utilization of the site resources by different species in
mixed stands can result in greater yields compared to pure
stands because interspecific competition becomes less intense
than intraspecific competition (Kelty and Cameron 1995).
This can occur under different mechanisms (competition
reduction and facilitation principles). A stratified canopy is
one of these mechanisms. Stratified species may experience
reduced competition by capturing light at different intensities
and locations within the canopy. If this happens, a stratified
canopy composed of shade-tolerant species growing underneath a shade-intolerant species would collectively intercept
© 2006 NRC Canada
1492
Can. J. For. Res. Vol. 36, 2006
Fig. 9. Relative yields of volume per hectare at age 12 for
Douglas-fir and western hemlock grown in mixed stands at three
density levels.
1.2
MIX
1.07
df/MIX
Relative Yield Total
1.0
wh/MIX
0.77
0.8
0.79
0.75
0.59
0.6
0.51
0.4
0.28
0.2
0.26
0.16
0.0
400
600
800
1000
1200
1400
1600
1800
Density (trees/ha)
more photosynthetically functional light than a pure stand of
the most shade-intolerant species and would maximize the
use of it (Kelty 1989). The shade-intolerant component of
the stand may, therefore, experience less intense interspecific
than intraspecific competition when growing in a pure stand.
Interspecific and intraspecific competition were assessed
in this study by changes in the dimensions of each tree species
growing in the mixture compared to their growth in the
monocultures (Menalled et al. 1998). It was found that
Douglas-fir trees experienced an increase in diameter,
height, and individual-tree volume in the mixture compared
to in the pure stand. In contrast, the opposite effect was
found for western hemlock, which experienced a growth
reduction due to the presence of Douglas-fir. The effect that
western hemlock has on Douglas-fir, seems minimal — growth
is similar to that with a wider spacing in pure Douglas-fir
stands. Unfortunately, vertical foliage profiles and measurements of light interception were not available to truly assess
this; however, the changes in crown size and position
(Amoroso 2004) further support the idea of stratification and
reduced competition experienced by the Douglas-fir component. This could result in reduced interspecific competition
in the mixed stand compared to intraspecific competition
experienced by Douglas-fir when growing in a pure stand.
Another possible way to examine the effects of interspecific
and intraspecific competition is through the height/diameter
ratio (h/d). This ratio has been used as a measure of competition in even-aged stands (Abetz 1976), because trees allocate
more carbon to height than to diameter growth to participate
in the canopy (Bauhus et al. 2000). Even though differences
were not statistically significant at the 95% confidence level,
the slightly lower h/d that Douglas-fir trees exhibited growing
in the mixture compared to that in the monoculture is consistent with Wang et al. (2000), who reported that intolerant
species growing in mixed stands allocate more carbon to
stemwood. In contrast to this, the higher h/d found for western
hemlock in the mixture compared to in the pure stand
suggests that interspecific competition is greater than
intraspecific competition. This could mean that western
hemlock needs to allocate relatively more resources to
height growth to participate in the canopy, while Douglas-fir
becomes situated in an upper position in the canopy earlier
and could allocate more resources to diameter growth.
Overall stand productivity: pure versus mixed stands
While in most of the common situations mixtures will not
exhibit higher yields than the most productive monoculture,
many studies have shown that both natural and planted
mixed stands can yield as much or more than pure stands of
the most productive of their components (Wierman and
Oliver 1979; Kelty 1989; Binkley 1992; Brown 1992;
Montagnini et al. 1995; 2000, DeBell et al. 1997; Man and
Lieffers 1999; DeBell et al. 1997; Khanna 1997; Bauhus et
al. 2000; Balieiro et al. 2002; Garber and Maguire 2004). In
some of these cases the increased productivity was the result
of species interactions typified by the facilitation production
principle, while others were typical of the competitive
production principle (Vandermeer 1989). The Douglas-fir –
western hemlock interactions examined in this study are
explained by the competitive production principle, where
species growing in a mixture utilize resources differently,
since resources unexploited by one species are used by the
other species.
The productivity among the pure and mixed stands was
compared both on a relative (relative yield, sensu Harper
1977) and absolute basis (volume per hectare). The relative
yield analysis revealed that at 1729 trees/ha the RYT of the
mixture was, although slightly higher, statistically indistinguishable from 1.0 (Fig. 9). In terms of absolute volume per
hectare, the mixed stand had the same yield as the monocultures. The increased yield observed in the mixed stand at
this density was probably driven by the differences of
photosynthetic efficiency of foliage and the partial stratification observed. The stratified canopy probably resulted in
sufficient radiation interception in the upper canopy to allow
higher productivity of the shade-intolerant species and yet
adequate transmission of radiation to the shade-tolerant
species in the lower position of the canopy (Menalled et al.
1998). This might have resulted in a maximization of light
use by the canopy due to increased light efficiency and
reduction in the competition for this resource (Kelty 1992).
As a result, it appears that the Douglas-fir component was
enhanced in the mixture, expressed by a greater mean
individual-tree volume, probably due to an increase in the
light interception. In contrast, the yield of western hemlock
was disproportionately low as compared with that of pure
western hemlock plantations.
The volume per hectare response for western hemlock
turns down slightly within the range of densities studied,
though not significantly so, but may still require some explanation. This response is not likely to be observed in actual
forest stands; there are no reports in the literature indicating
such a reduction in productivity for western hemlock at
higher densities. It is our assumption that this is a modeling
artifact caused in part by higher densities for pure western
hemlock being less well represented in our data than in the
other species mixes, making the western hemlock model less
© 2006 NRC Canada
Amoroso and Turnblom
1493
Fig. 10. Absolute and relative yield for the Douglas-fir (DF) and western hemlock (WH) monocultures and the 50/50 mixture at three
density levels: 494, 1111, and 1729 trees/ha. Note that species composition is labeled on the horizontal axis as % DF / % WH.
494 trees/ha
45
494 trees/ha
1.2
40
1
Relative Yield
3
Yield (m /ha)
35
30
25
20
15
10
0.8
0.6
0.4
0.2
5
0
0/100
50/50
0
0/100
100/0
50/50
1111 trees
90
100/0
1111 trees
1.2
80
1
60
Relative Yield
3
Yield (m /ha)
70
50
40
30
20
0.8
0.6
0.4
0.2
10
0
0/100
50/50
0
0/100
100/0
1729 trees
120
1.2
105
50/50
100/0
1729 trees/ha
1
Relative Yield
75
3
Yield (m /ha)
90
60
45
0.8
0.6
0.4
30
0.2
15
0
0/100
50/50
100/0
Species % in the Stand
DF
WH
0
0/100
50/50
100/0
Species % in the Stand
MIX
precise in this range. This is one of the disadvantages of
using a quadratic model to represent nonlinearity — the
model can peak (or trough) illogically within the range of
the data. The downturn probably exacerbated by backtransforming (exponentiating) the fitted volume per hectare
response to the original scale. However, based on the observed
trend prior to this downturn, there certainly is no evidence
suggesting that pure western hemlock stands would have
statistically significantly higher volume per hectare than
DF
WH
MIX
both the pure Douglas-fir and mixed stands. In other words,
given a model that behaved more logically within the range
of the densities examined, the results of the study would be
unchanged.
Density and its effects on species interactions and
productivity
The effects of density on tree and stand growth and yield
in pure stands are well known; however, species growing in
© 2006 NRC Canada
1494
mixed stands would probably have different behavior at different densities compared to pure stands (Kelty and
Cameron 1995).
Most studies comparing mixtures and monocultures were
conducted at a single density. Densities utilized were usually
high, and even though it is not clear why these densities
were chosen, one could expect that high-density stands were
chosen to accelerate interactions among individuals and to
produce results in a shorter period of time. In an experiment
using two densities, Khanna (1997) observed that interactions between the species occurred earlier in the highdensity treatment, and they became evident later at the lowdensity level. Khanna also suggested that positive interactions between species might occur later in wider spacings.
Working with Douglas-fir and red alder (Alnus rubra) seedlings, Shainsky and Radosevich (1992) demonstrated that simultaneous manipulation of the densities of the two species
produced quantitative changes not only in tree growth, but
also in light, soil moisture, and leaf water potential. Additionally, Garber and Maguire (2004), working with two
mixed-species spacing trials using a range of densities similar to that used in this study, concluded that spacing is an
important factor determining species interactions that affect
relative performance of individual species.
The study presented here demonstrates the important role
density plays in the development and productivity of mixed
stands when compared to pure stands of its components. It
appears that interactions between the species involved in a
mixture occur in different degrees depending on the amount
of resources they are obligated to share and (or) for which
they compete. Relative yield results at 494 trees/ha show
that the RY for Douglas-fir was as expected (0.5) but, on the
other hand, western hemlock had an RY lower than expected; together, they had a combined RYT value of 0.77. At
1111 trees/ha, Douglas-fir RY was slightly higher than expected, while western hemlock RY and RYT still remained
below expectation but greater than observed at 494 trees/ha.
However, at 1729 trees/ha the combined RYT was 1.07 and
the Douglas-fir RY was much higher than expected. Although a potential productivity advantage for the mixture
may exist at 1729 trees/ha, this phenomenon has not yet appeared at the lower densities; western hemlock growth is adversely affected, and beneficial growth of Douglas-fir is only
starting to emerge at 1111 trees/ha at age 12. Continued
monitoring in the future is needed to see whether and when
these interactions do occur at lower densities.
Even supposing that the results found at 1729 trees/ha
would appear in the future at 1111 trees/ha, it is not likely
that the 494 trees/ha stands will respond in the same way.
Kelty and Cameron (1995) suggest that the use of low densities in mixed stands at establishment may allow species with
a slow juvenile growth rate to escape early suppression, and
as a consequence of this the species may not be able to express differences in their ability to use resources (i.e., stratification).
In addition to the result that mixed Douglas-fir – western
hemlock planted stands were able to produce as much volume per hectare as the pure stands of either species, mixed
stands may produce other economic benefits, though not
considered in this study. For example, wood quality of the
Douglas-fir trees growing in mixture may be improved, and
Can. J. For. Res. Vol. 36, 2006
this will be represented not only by the greater size of the
crop trees but also by the improvement in the bole quality as
a result of western hemlock trees shading the lower limbs.
This might include natural pruning as well as a reduction in
limb diameter (knots). Along with the timber production
objectives, vertical complexity in mixed stands can also
provide other benefits such as a broader range in wildlife
habitats and enhanced aesthetic values without the often
unsubstantiated timber yield trade-off. It has also been
suggested that mixed plantations of these two species may
be more wind firm and disease resistant than pure stands
(Wierman and Oliver 1979).
Conclusions
It appears that interactions between species growing in a
mixed stand occur to different degrees depending on the
amount of resources they are obligated to share and (or) for
which they compete. The study presented here supports the
important role density plays on the productivity of mixed
stands.
In terms of ecological theory, it appears that “the competitive production principle” has contributed to superior yields
of the mixed planted stands at high densities, achieving as
much productivity as the pure stands. This was a result of
the partial stratification observed and the presumably better
use of the site resources made by the two species in the mixture compared to in the pure stands. Because of this, less
interspecific competition in the mixed stands than
intraspecific competition in the pure stands was probably
experienced. At low and medium densities, however, interactions between species did occur (as deduced from dimensional changes), but may not have been of great enough
magnitude to cause the mixture to outperform the pure plantations in terms of total yield.
The results presented here reflect responses at an early
stage in stand development. Long-term measurements are
expected to show other effects, probably making the interactions among species more evident. In addition, comparisons
among single- and mixed-species plantations are few, and
the advantages and disadvantages may be site specific.
Mixtures that could achieve a reduction in interspecific competition might, as a consequence of this, increase their yields
over the pure stands of its components. However, this is
likely to happen only if the supply of the resource for which
competition is reduced is limiting the production in the pure
stands (Kelty 1992). For these reasons the results of this
study should be understood locally and not be extrapolated
too widely to other sites.
Acknowledgements
This research was supported by the Stand Management
Cooperative, College of Forest Resources, University of Washington, Seattle, Washington, and was part of an M.Sc. thesis
submitted at the same university. The authors wish to thank
Bob Gonyea and Bert Hasselberg for their suggestions and
help with the data collection, Randy Collier for his assistance with the database and analyses, and Dr. Bruce Larson
and Dr. David Briggs for their comments on earlier versions
© 2006 NRC Canada
Amoroso and Turnblom
of the manuscript. We also appreciate the valuable comments of two anonymous reviewers.
References
Abetz, P. 1976. Beiträge zum Baumwachstum. Der h/d-Wert –
Mehr als ein Schlankheitsgrad. Forst Holzwirt. 31: 389–393. [In
German.]
Amoroso, M.M. 2004. Are mixed species stands more productive
than single species stands? Douglas-fir and western hemlock
plantations in the Pacific Northwest. M.Sc. thesis, College of
Forest Resources, University of Washington, Seattle, Wash.
Assmann, E. 1970. The principles of forest yield study. Pergamon
Press, Oxford.
Balieiro, F.C., Fontes, R.L.F., Dias, L.E., Franco, A.A., Campello,
E.F.C., and de Faria, S.M. 2002. Accumulation and distribution
of aboveground biomass and nutrients in pure and mixed stands
of guachapele and eucalyptus. J. Plant Nutr. 25(12): 2639–2654.
Bauhus, J., Khanna, P.K., and Menden, N. 2000. Aboveground and
belowground interactions in mixed plantations of Eucalyptus
globulus and Acacia mearnsii. Can. J. For. Res. 30: 1886–1894.
Binkley, D. 1983. Ecosystem production in Douglas-fir plantations:
interaction of red alder and site fertility. For. Ecol. Manage. 5:
215–227.
Binkley, D. 1992. Mixtures of nitrogen-fixing and non-nitrogenfixing tree species. In The ecology of mixed species stands of
trees. Edited by M.G.R. Cannell, D.C. Malcolm, and P.A. Robertson. Blackwell Scientific, London. pp. 99–113.
Brown, A.H.F. 1992. Functioning of mixed-species stands at
Gisburn, N.W. England. In The ecology of mixed species stands
of trees. Edited by M.G.R. Cannell, D.C. Malcolm, and P.A.
Robertson. Blackwell Scientific, London. pp. 99–113.
Bruce, D., and DeMars, D.J. 1974. Volume equations for secondgrowth Douglas-fir (Pseudotsuga menziesii). USDA For. Serv.
Res. Note PNW-239.
Crowder, M.J., and Hand, D.J. 1990. Analysis of repeated measures. Chapman and Hall. Monogr. Stat. Appl. Prob. 41.
DeBell, D.S., Cole, T.G., and Whitesell, C.D. 1997. Growth, development and yield in pure and mixed stands of Eucalyptus and
Albizia. For. Sci. 43: 286–298.
Flewelling, J.W., and Raynes, L.M. 1993. Variable-shape stem profile predictions for western hemlock. Part I. Predictions from
DBH and total height. Can. J. For. Res. 23: 520–536.
Garber, S.M., and Maguire, D.A. 2004. Stand productivity and
development in two mixed-species spacing trials in the Central
Oregon Cascades. For. Sci. 50(1): 92–105.
Halloin, L.J. 1987. Soil survey of Clallam County area, Washington.
USDA Soil Conservation Service, Washington, D.C., in cooperation with Washington State Department of Natural Resources,
Washington State University, Agriculture Research Center, and
Clallam County Commissioners.
Harper, J.L. 1977. Population biology of plants. Academic Press,
London.
Kelty, M.J. 1989. Productivity of New England hemlock/hardwood
stands as affected by species composition and canopy structure.
For. Ecol. Manage. 28: 237–257.
Kelty, M.J. 1992. Comparative productivity of monocultures and
mixed species stands. In The ecology and silviculture of mixedspecies forests. Edited by M.J. Kelty, B.C. Larson, and C.D.
Oliver. Kluwer Academic Publishers, Dordrecht, Netherlands.
pp. 125–141.
Kelty, M.J., and Cameron, I.R. 1995. Plot designs for the analysis
of species interactions in mixed stands. Commonw. For. Rev.
74(4): 322–332.
1495
Khanna, P.K. 1997. Comparison of growth and nutrition of young
monocultures and mixed stands of Eucalyptus globulus and
Acacia mearnsii. For. Ecol. Manage. 94: 105–113.
King, J.E. 1958. Development of a stand of coniferous reproduction and interplanted Douglas-fir. Northwest Sci. 32(1): 1–8.
King, J.E. 1966. Site index curves for Douglas-fir in the Pacific
Northwest. Weyerhaeuser For. Pap. 8.
Larson, B.C. 1986. Development and growth of even-aged stands
of Douglas-fir and grand fir. Can. J. For. Res. 16: 367–372.
Lewis, J.D., McKane, R.B., Tingey D.T., and Beedlow, P.A. 2000.
Vertical gradients in photosynthetic light response within an oldgrowth Douglas-fir and western hemlock canopy. Tree Physiol.
20: 447–456.
Maguire, D.A., Bennett, W.S., Kershaw, J.A., Jr., Gonyea, R., and
Chappell, H.N. 1991. Establishment report [unpublished report].
Stand Management Cooperative Silviculture Project field installations. Stand Management Cooperative, College of Forest
Resources, University of Washinton, Box 352100, Seattle, WA
98195-2100.
Man, R., and Lieffers, V.J. 1999. Are mixtures of aspen and white
spruce more productive than single species stands? For. Chron.
75(3): 505–512.
Martin, F.C., and Flewelling, J.W. 1998. Evaluation of tree height
prediction models for stand inventory. West. J. Appl. For. 13(4):
109–119.
Menalled, F.D., Kelty, M.J., and Ewel, J.J. 1998. Canopy development in tropical tree plantations: a comparison of species mixtures and monocultures. For. Ecol. Manage. 104: 249–263.
Montagnini, F. 2000. Accumulation in above-ground biomass and
soil storage of mineral nutrients in pure and mixed plantations in
a humid tropical land. For. Ecol. Manage. 134: 257–270.
Montagnini, F., Gonzalez, E., Porras, C., and Reingans, R. 1995.
Mixed and pure forest plantations in the humid neotropics: a
comparison of early growth, pest damage and establishment costs.
Commonw. For. Rev. 74(4): 306–314.
Oehlert, G.W. 2000. A first course in design and analysis of experiments. W.H. Freeman and Company, New York.
Oliver, C.D., and Larson, B.C. 1996. Forest stand dynamics.
McGraw-Hill, New York.
Pringle, R.F. 1986. Soil survey of Grays Harbor County Area,
Pacific County, and Wahkiakum County, Washington. USDA
Soil Conservation Service, Washington, D.C., in cooperation
with Washington State Department of Natural Resources, Washington State University, and Agriculture Research Center.
SAS Institute Inc. 1999–2001. SAS for Windows version 8. SAS
Institute Inc., Cary, N.C.
Scholz, D.M., and Smith, J.H.G. 1975. Comparison of crown
dimensions of open grown Douglas-fir, western hemlock and
western red cedar. Reforestation Notes, April 1975. Reforestation Division, British Columbia Forest Service, Victoria, B.C.
Shainsky, L.J., and Radosevich, S.R. 1992. Mechanisms of competition between Douglas-fir and red alder seedlings. Ecology,
73(1): 30–45.
Smith, D.M., Larson, B.C., Kelty M.J., and Ashton, P.M.S. 1997.
The practice of silviculture: applied forest ecology. John Wiley
& Sons, New York.
Vandermeer, J.H. 1989. The ecology of intercropping. Cambridge
University Press, Cambridge.
Wang, J.R., Letchford, T., Comeau, P., and Kimmins, J.P. 2000.
Above- and below-ground biomass nutrient distribution of a paper
birch and subalpine fir mixed-stand in the Sub-Boreal Spruce
zone of British Columbia. For. Ecol. Manage. 130: 17–26.
Western Regional Climate Center. 2003. Western U.S. climate
historical summaries [online]. Western Regional Climate Center,
© 2006 NRC Canada
1496
Reno, Nev. Available from http://www.wrcc.dri.edu/CLIMATEDATA.
html [accessed 15 January 2003].
Wierman, C.A., and Oliver, C.D. 1979. Crown stratification by
species in even-aged mixed stands of Douglas-fir – western
hemlock. Can. J. For. Res. 9: 1–9.
Appendix A
The following standard response surface model was fit to
all growth and yield variables:
Y = b0 + b1 (c1) + b2 (c2) + b3 (sp1) + b4 (sp2) + b5 (TPH) +
b6 (age) + b7 (c1 × TPH) + b8 (c2 × TPH) + b9 (sp1 × TPH)
+ b10 (sp2 × TPH) + b11 (c1 × age) + b12 (c2 × age) + b13
(sp1 × age) + b14 (sp2 × age) + b15 (TPH × age) + b16 (TPH ×
TPH) + b17 (age × age) + b18 (c1 × TPH × age) + b19 (c2 ×
TPH × age) + b20 (sp1 × TPH × age) + b21 (sp2 × TPH ×
age) + b22 (c1 × TPH × TPH) + b23 (c2 × TPH × TPH) + b24
(sp1 × TPH × TPH) + b25 (sp2 × TPH × TPH) + b26 (c1 ×
age × age) + b27 (c2 × age × age) + b28 (sp1 × age × age) +
b29 (sp2 × age × age)
Can. J. For. Res. Vol. 36, 2006
where
Y is a transformation of response variables (square-root
for mean DBH; logarithm for mean total height, mean individual tree volume, and volume per hectare);
c1 is a categorical variable indicating Douglas-fir as a
component of the mixed stands (1 is Douglas-fir in mixture,
0 otherwise);
c2 is a categorical variable indicating western hemlock as
a component of the mixed stands (1 is western hemlock in
mixture, 0 otherwise);
sp1 is a categorical variable indicating Douglas-fir as the
single component in the pure stands (1 is pure Douglas-fir
stands, 0 otherwise);
sp2 is a categorical variable used to indicate western hemlock as the single component in the pure stands (1 is pure
western hemlock stands, 0 otherwise);
TPH is density expressed in number of trees per hectare;
age is plantation age in years;
b0 to b29 are ordinary least squares parameter estimates for
each variable and interaction term.
© 2006 NRC Canada